标题： 一种用于高采样率光学巡天瞬变检测的新方法：应用于M31新星的系统性搜索 显示英文标题

标题： A novel method for transient detection in high-cadence optical surveys: Its application for a systematic search for novae in M31

摘要： 在大规模时域巡天中，从数据获取到源探测的整个处理过程通常是自动化的。 其中一个主要挑战是受伪迹污染的影响，特别是在强未分辨发射区域尤为显著。 我们提出了一种新颖的方法，用于从这类巡天数据流水线的输出中识别变源和暂现源候选体。 我们将该方法应用于iPTF对M31星系核区域观测数据的系统性新星搜寻。 我们证明了iPTF流水线产生的大多数伪迹形成了一个局部均匀的虚假探测背景，大致服从泊松统计分布， 而真实的变源和暂现源以及与亮星相关的伪迹则导致了探测簇集，其扩散范围由源定位精度决定。 这使得问题类似于利用X射线望远镜生成图像上的源探测，从而可以利用X射线天文学开发的工具。 特别是，我们使用了来自Chandra数据分析包CIAO中的基于小波的源探测算法。 从iPTF数据流水线生成的约2.5×10⁵个原始探测中，我们得到了约4000个独特的源候选体。 通过将这些候选体与深参考图像的源目录进行交叉匹配，我们找到了其中约90%的对应体。 这些对应体要么是由于点扩散函数匹配不完美造成的伪迹，要么是真实的变源。 剩下的约400个探测是暂现源。 我们通过对新星光变曲线的预期性质应用选择标准，从这些候选体中识别出新星。 因此，我们找回了巡天期间记录的所有12个已知新星，并发现了三个新星候选体。 我们的方法具有通用性，可用于从光学时域数据的伪迹中挖掘任何目标。 由于它是完全自动化的，其不完整性可以被准确计算并加以修正。 显示英文摘要

摘要： [abridged] In large-scale time-domain surveys, the processing of data, from procurement up to the detection of sources, is generally automated. One of the main challenges is contamination by artifacts, especially in regions of strong unresolved emission. We present a novel method for identifying candidates for variables and transients from the outputs of such surveys' data pipelines. We use the method to systematically search for novae in iPTF observations of the bulge of M31. We demonstrate that most artifacts produced by the iPTF pipeline form a locally uniform background of false detections approximately obeying Poissonian statistics, whereas genuine variables and transients as well as artifacts associated with bright stars result in clusters of detections, whose spread is determined by the source localization accuracy. This makes the problem analogous to source detection on images produced by X-ray telescopes, enabling one to utilize tools developed in X-ray astronomy. In particular, we use a wavelet-based source detection algorithm from the Chandra data analysis package CIAO. Starting from ~2.5x10^5 raw detections made by the iPTF data pipeline, we obtain ~4000 unique source candidates. Cross-matching these candidates with the source-catalog of a deep reference image, we find counterparts for ~90% of them. These are either artifacts due to imperfect PSF matching or genuine variable sources. The remaining ~400 detections are transient sources. We identify novae among these candidates by applying selection cuts based on the expected properties of nova lightcurves. Thus, we recovered all 12 known novae registered during the time span of the survey and discovered three nova candidates. Our method is generic and can be applied for mining any target out of the artifacts in optical time-domain data. As it is fully automated, its incompleteness can be accurately computed and corrected for.